Project Summary Mathematical ecology models of host-microbiota interaction in auto microbiota transplants (auto-FMT) We aim to develop mathematical models for the rational design of microbiota transplants that can restore compositional diversity and function to the damaged microbiota of antibiotic-treated patients. We will focus on hospitalized cancer patients receiving allogeneic hematopoietic stem cell transplants (allo-HSCT). Allo-HSCT is a potentially curative cancer treatment that compromises the immune system, and requires that patients receive massive antibiotic treatments to prevent and treat life-threatening infections. We will build on a vast clinical database, in vitro experiments in bioreactors and in vivo experiments with mice to develop dynamic mathematical models that describe how antibiotics cause changes in the microbial composition, and how that can impact the recovery of the host's immune system after allo-HSCT. The model expands approaches pioneered by our team?the Generalized Lotka Volterra Ecological Regression (GLOVER) and agent-based models?towards a model that can assist in the development of microbiota therapies for patients undergoing allo-HSCT. In aim 1 we will use data from a unique clinical resource available at the Memorial Sloan Kettering Cancer center?a sample bank obtained from >1,500 allo-HSCT patients (including microbiome 16S rRNA and shotgun sequencing) and extensive clinical metadata (including time series of complete blood counts and time and doses of all drugs given while the patients are hospitalized); we will also use data from a first-of-its-kind controlled randomized trial of autologous fecal microbiota transplant (auto-FMT) undergoing in allo-HSCT patients. We will use these unique resources to parameterize our models and investigate how the microbiota composition influences the recovery of the host immune system. In aim 2 we will validate the microbial component of our mathematical model using experimental data from anaerobic laboratory reactors that recreate?in vitro?the human microbiota dynamics during antibiotic treatment and auto-FMT in the absence of a living host. In aim 3 we will develop mouse models to investigate those same microbiota dynamics experimentally but now in the context of a living host. The data obtained from these clinical studies, in vitro experiments and in vivo models will refine our mathematical models in close cycles of simulation and quantitative experimentation. Our ultimate goal is to develop models that can define optimal microbial cocktails and reconstitute the perturbed microbiota of allo- HSCT patients. In the process we hope to uncover general principles of microbiota ecology for future therapies in other patient populations whose microbiota is damaged by antibiotic treatments.